Improved Human Tracking Algorithm Based on Color Histogram

Chen Chunyu,Yao Qiuxiang,Qiao Yulong
2012-01-01
Abstract:A new human tracking algorithm combining the color histogram and the improved model update algorithm is proposed.Firstly,background subtraction is used to detect all candidate moving models of surveillance area,and then in the HSV color space,the color components of target and candidate models are quantified with unequal interval to obtain characteristic vectors.Through the similarity calculation,the current target model is replaced by the optimal similar model,thus a multi-target tracking is achieved.Meanwhile,under occlusion,the model update algorithm is improved to accurately track moving targets.Experimental results show that the algorithm is computationally efficient and robust to rotation,scale invariant and interactions of non-rigid objects.
What problem does this paper attempt to address?